CN109886233B - Aquaculture state monitoring method and device based on hidden Markov model - Google Patents
Aquaculture state monitoring method and device based on hidden Markov model Download PDFInfo
- Publication number
- CN109886233B CN109886233B CN201910153100.0A CN201910153100A CN109886233B CN 109886233 B CN109886233 B CN 109886233B CN 201910153100 A CN201910153100 A CN 201910153100A CN 109886233 B CN109886233 B CN 109886233B
- Authority
- CN
- China
- Prior art keywords
- aquaculture
- fish
- culture
- state
- shoal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/80—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in fisheries management
- Y02A40/81—Aquaculture, e.g. of fish
Landscapes
- Farming Of Fish And Shellfish (AREA)
Abstract
The invention discloses an aquaculture state monitoring method based on a hidden Markov model, which comprises the following steps: s1: establishing an aquaculture model based on a hidden Markov model according to the relationship between hidden parameters extracted by the culture rule and the culture dominant state, wherein the hidden parameters are used for representing the information of the culture abnormal state; s2: acquiring culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information; s3: obtaining an observation sequence according to the dominant state of cultivation; s4: and calculating various abnormal probabilities according to the observation sequence and the aquaculture model, if the abnormal probabilities are larger than the corresponding abnormal thresholds, considering that the corresponding abnormality occurs, and if not, considering that the abnormality does not occur. The invention has the technical characteristics of comprehensive monitoring, real-time monitoring and accurate abnormality monitoring.
Description
Technical Field
The invention belongs to the technical field of aquaculture, and particularly relates to an aquaculture state monitoring method and device based on a hidden Markov model.
Background
At present, with the development of the technology of the Internet of things, aquaculture is more and more modern. The application of various sensing technologies facilitates the real-time monitoring of aquaculture. The existing aquaculture technology can mainly realize remote monitoring and remote control feeding, and the process still needs manual intervention.
The prior art can mainly monitor the culture environment and analyze the state of fish shoal. However, the whole process mainly focuses on the monitoring of dominant parameters, and implicit factors such as hunger, epidemic disease, silt, algae and the like are not considered, are not directly equivalent to the directly monitored physical parameters, and the corresponding relationship between the implicit factors and the dominant factors is not a simple mapping relationship.
Disclosure of Invention
The invention aims to provide an aquaculture state monitoring method and device based on a hidden Markov model, which have the technical characteristics of comprehensive monitoring, real-time monitoring and accurate abnormal monitoring.
In order to solve the problems, the technical scheme of the invention is as follows:
an aquaculture state monitoring method based on a hidden Markov model comprises the following steps:
s1: establishing an aquaculture model based on a hidden Markov model according to the relationship between hidden parameters extracted by the culture rule and the culture dominant state, wherein the hidden parameters are used for representing the information of the culture abnormal state;
s2: acquiring the culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information;
s3: obtaining an observation sequence according to the culture dominant state;
s4: and calculating various abnormal probabilities according to the observation sequence and the aquaculture model, if the abnormal probabilities are larger than corresponding abnormal thresholds, the corresponding abnormality is considered to occur, and if not, the abnormality is considered not to occur.
According to an embodiment of the present invention, the step S1 specifically includes the following steps:
constructing a fish swarm dominant state matrix according to the samples of the aquaculture water environment information and the fish swarm activity information at a plurality of time points, and simultaneously recording the fish swarm state at each time point to form a relationship between the recessive parameters and the aquaculture dominant state;
performing hidden Markov model training according to the shoal state and the shoal dominant state matrix to obtain the optimal aquaculture model;
and calculating the parameters of the aquaculture model by adopting a Baum-Welch algorithm, and verifying the training result of the aquaculture model.
According to an embodiment of the present invention, in the step S2, the step of obtaining the fish school activity information includes:
reading and scanning culture monitoring pictures at different moments, comparing the culture monitoring pictures with the fish swarm characteristic values, and obtaining fish swarm position information matrixes at different moments;
and calculating the shoal of fish activity information according to the shoal of fish position information matrixes at different moments, wherein the shoal of fish activity information comprises the speed, the acceleration, the depth and the dispersion of the shoal of fish.
According to an embodiment of the present invention, the step S3 specifically includes the following steps:
and combining the aquaculture water environment information and the fish school activity information at each same time point according to the aquaculture water environment information and the fish school activity information to form the time-based observation sequence.
According to an embodiment of the present invention, the step S4 further includes a step S5:
if the abnormality occurs, the cultivation measure corresponding to the abnormality is executed to correct the abnormality.
An aquaculture state monitoring device based on a hidden markov model, comprising:
the model building module is configured to build an aquaculture model based on a hidden Markov model according to the relationship between the hidden parameters extracted by the culture rule and the culture dominant state, wherein the hidden parameters are used for representing the culture abnormal state information;
the detection sensing module is configured to acquire the culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information;
the data processing module is configured to acquire an observation sequence according to the culture dominant state;
the data processing module is further configured to calculate various abnormal probabilities according to the observation sequence and the aquaculture model, if the abnormal probabilities are larger than corresponding abnormal thresholds, the corresponding abnormality is considered to occur, and if not, the abnormality is considered not to occur.
According to an embodiment of the present invention, the model building module includes a sample unit, a training unit, and a verification unit;
the sample unit is configured to construct a fish swarm dominant state matrix according to samples of the aquaculture water environment information and the fish swarm activity information at a plurality of time points, and record fish swarm states at each time point at the same time so as to form a relationship between recessive parameters and the aquaculture dominant states;
the training unit is configured to perform hidden Markov model training according to the shoal state and the shoal dominant state matrix so as to obtain the optimal aquaculture model;
the verification unit is configured to calculate the aquaculture model parameters by using a Baum-Welch algorithm and verify training results of the aquaculture model.
According to an embodiment of the invention, the detection sensing module comprises a receiving and processing unit, a water environment sensor and a vision sensor;
the receiving and processing unit is configured to receive the water environment data monitored by the water environment sensor to form the aquaculture water environment information;
the receiving processing unit is further configured to read and scan the culture monitoring pictures at different moments monitored by the visual sensor, compare the culture monitoring pictures with the fish school characteristic values and acquire fish school position information matrixes at different moments;
the receiving processing unit is further configured to calculate the shoal of fish activity information according to the shoal of fish position information matrixes at different moments, wherein the shoal of fish activity information comprises a shoal of fish speed, acceleration, depth and dispersion.
According to an embodiment of the invention, the data processing module is specifically configured to combine the aquaculture water environment information and the fish school activity information at each same point in time according to the aquaculture water environment information and the fish school activity information to form the time-based observation sequence.
According to an embodiment of the present invention, the system further includes an exception handling module configured to execute a cultivation measure corresponding to the exception to perform exception correction.
By adopting the technical scheme, the invention has the following advantages and positive effects compared with the prior art:
compared with common aquaculture monitoring, the method provided by the invention has the advantages that the hidden parameters related to aquaculture and the relation between the hidden parameters and the dominant parameters are considered, the defect that the monitoring parameters are only the dominant parameters in the traditional method is overcome, meanwhile, for the prediction of abnormal states, the prediction and control are more intelligent and scientific through multi-parameter monitoring of the hidden parameters and the dominant parameters, and the technical effects of comprehensive monitoring, real-time monitoring and accurate abnormal monitoring are achieved.
Drawings
FIG. 1 is a schematic flow chart of an aquaculture state monitoring method based on a hidden Markov model;
FIG. 2 is a schematic diagram of a hidden Markov model training process of an aquaculture state monitoring method based on a hidden Markov model according to the present invention;
fig. 3 is a schematic diagram of an aquaculture state monitoring method based on a hidden markov model according to the present invention.
Detailed Description
The invention provides a hidden Markov model-based aquaculture state monitoring method and device, which are further described in detail below with reference to the accompanying drawings and specific embodiments. Advantages and features of the invention will become more apparent from the following description and from the claims.
Example 1
Referring to fig. 1, the present embodiment provides a method for monitoring aquaculture conditions based on a hidden markov model, comprising the steps of:
s1: establishing an aquaculture model based on a hidden Markov model according to the relationship between hidden parameters extracted by the culture rule and the culture dominant state, wherein the hidden parameters are used for representing the information of the culture abnormal state;
specifically, the step S1 specifically includes the steps of:
constructing a fish swarm dominant state matrix according to samples of the aquaculture water environment information and the fish swarm activity information at a plurality of time points, and simultaneously recording the fish swarm states at each time point to form a relationship between recessive parameters and aquaculture dominant states;
according to the fish swarm states and the fish swarm dominant state matrix, performing hidden Markov model training to obtain an optimal aquaculture model;
calculating parameters of the aquaculture model by adopting a Baum-Welch algorithm, and verifying training results of the aquaculture model;
s2: acquiring culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information;
specifically, the step of obtaining the fish school activity information comprises the following steps:
reading and scanning breeding monitoring pictures at different moments, comparing the breeding monitoring pictures with fish swarm characteristic values, and obtaining fish swarm position information matrixes at different moments;
according to the shoal position information matrixes at different moments, calculating to obtain shoal activity information, wherein the shoal activity information comprises shoal speed, acceleration, depth and dispersion;
s3: obtaining an observation sequence according to the dominant state of cultivation;
specifically, the step S3 specifically includes the steps of:
combining the aquaculture water environment information and the fish school activity information at the same time point according to the aquaculture water environment information and the fish school activity information to form a time-based observation sequence;
s4: and calculating various abnormal probabilities according to the observation sequence and the aquaculture model, if the abnormal probabilities are larger than the corresponding abnormal thresholds, considering that the corresponding abnormality occurs, and if not, considering that the abnormality does not occur.
Preferably, step S4 is followed by step S5:
if the abnormality occurs, the cultivation measure corresponding to the abnormality is executed to correct the abnormality.
The present embodiment will now be described in detail:
1) S1: according to the relation between the hidden parameters extracted by the culture rule and the culture dominant state, establishing an aquaculture model based on a hidden Markov model, wherein the hidden parameters are used for representing the culture abnormal state information:
according to the cultivation law, common recessive parameters are extracted, wherein the common recessive parameters comprise abnormal states such as oxygen deficiency of the fish shoal, hunger of the fish shoal, illness of the fish shoal, excessive sludge, excessive algae, excessive garbage and the like, a hidden Markov model is built aiming at each abnormal state, the extraction is mainly carried out according to a investigation method, daily cultivation processes are observed, questionnaires are carried out on cultivation staff, the common abnormal states are extracted, and when the abnormal states occur, the current monitoring parameters are recorded.
According to the embodiment, the hidden parameters and the relation between the hidden parameters and the dominant parameters are extracted from the culture rules, so that the technical effect of more comprehensive and scientific monitoring is achieved.
Referring to fig. 2, taking a fish-shoal anoxia abnormality as an example, using step S1 to obtain n fish samples, modeling sample water environment information PH, OX, T of m time points, and fish-shoal activity characteristics V, a, cx, cy, constructing a fish-shoal dominant state matrix O, and simultaneously recording a fish-shoal anoxia state q= [ Q1, Q2 … qT ], wherein 1 represents anoxia, 0 represents no anoxia, and performing hidden markov model training according to the anoxia abnormal state Q and the culture monitored dominant state matrix O, respectively, to obtain an optimal aquaculture model λ= (pi, a, B), wherein pi is an initial probability distribution of the anoxia state; a is a transfer matrix of a fish school in an anoxic state; b is the distribution probability of the display state.
Initializing a model, calculating model parameters by adopting a Baum-Welch algorithm, and when the error of the previous and the later iterations is smaller than a set threshold D, considering that the model training is finished, wherein the model training is specifically as follows:
the variable ζ (i, j) =p (qt=i, qt+1=j|o, λ) is defined, representing the probability of the anoxic state i at time t, the anoxic state j at time t+1.
Definition variable ζ (i) = Σ j ξ t And (i, j) represents the probability of being in the anoxic state i at the moment t under a given model and observation sequence.
The model parameters are estimated in the following way:
π＝ζ 1 (i) Representing the expected number of times in the anoxic state i when t=1;
The initial lambda= (pi, a, B) is chosen and substituted into the above formula to obtain a new set of lambda 1= (pi 1, a1, B1)
Calculating the |log (P (O|λ1)) -log (P (O|λ1))| when the value is smaller than the threshold D, and obtaining a final model. And when the value is larger than D, substituting lambda 1 into the formula, and repeating the steps until the error is smaller than D.
According to the method, the aquaculture model based on the hidden Markov model is built based on actual sample data, wherein the aquaculture model is trained by the hidden Markov model and verified by a Baum-Welch algorithm, so that the aquaculture model is more close to the actual aquaculture condition, and the technical effect of accurate anomaly monitoring is achieved.
2) Step S2: obtaining a culture dominant state, wherein the culture dominant state comprises culture water environment information and fish school activity information:
acquiring aquaculture water environment information:
NB DTU is connected with sensor ZZ-PHB-300 through first 485 communication interface to gather the pH value PH of aquaculture water, is connected with sensor ZZ-DOS-600 through second 485 communication interface to gather the dissolved oxygen content OX of aquaculture water, is connected with BPHT-V10 through third 485 communication interface to gather temperature information T of aquaculture water. Specifically, this embodiment illustrates three parameters, and in practice, the water environment information may include other parameters related to the aquaculture water environment.
Acquiring fish swarm activity information:
pre-storing a characteristic value S of the fish; reading a video monitoring picture, scanning an image, comparing with a prestored fish swarm characteristic value S, if the comparison is successful, identifying the fish as the fish, recording the position information (x 11, y 11) of the fish, scanning the whole image in sequence, and obtaining a position information matrix of the fish swarm, wherein when n fish are present, the position matrix is in the following format: (x 11, y11; x21, y21; … … xn1, yn 1);
and (5) at interval time t, reading a video monitoring picture at the next moment, and repeating the steps to obtain a fish school position matrix at the next moment: (x 12, y12; x22, y22; … … xn2, yn 2);
and then reading a video monitoring picture at the next moment at intervals of time t, repeating the steps, and obtaining a fish school position matrix at the next moment: (x 13, y13; x23, y23; … … xn3, yn 3);
according to the two moment shoal position matrixes, indexes representing the shoal activity information such as the shoal speed, the acceleration and the dispersion are calculated and obtained.
Taking the first fish as an example:
speed of:
time period 0-t:
t-2 t time period:
time period of 0 to 2 t:
acceleration:
time period of 0-2 t:
dispersion degree:
can be represented by a discrete coefficient at that instant, taking the first acquired image as an example:
wherein, the liquid crystal display device comprises a liquid crystal display device,
therefore, the dispersion is:
according to the method, the device and the system, the water environment information and the fish school activity information are monitored to obtain dominant parameters, so that the aquaculture water environment is monitored comprehensively, and the technical effect of real-time monitoring is achieved.
3) S3: according to the dominant state of cultivation, obtaining an observation sequence:
the method comprises the steps of monitoring aquaculture water environment information and fish swarm activity information in real time to obtain a time-based observation sequence O1, wherein the observation sequence further comprises water environment information PH1, OX1 and T1 and fish swarm activity characteristics V1, a1, cx1 and cy1 at each time point.
4) S4: according to the observation sequence and the aquaculture model, calculating to obtain various abnormal probabilities, if the abnormal probabilities are larger than corresponding abnormal thresholds, considering that corresponding abnormalities occur, and if not, considering that no abnormalities occur:
for the aquaculture model obtained in the above observation sequence O1 and step S2, the following formula is adopted:
let i=0, j=1 estimate the hidden state with the highest probability, taking the abnormal oxygen deficiency of the fish shoal as an example, the probability that the fish shoal is not oxygen deficient at the moment t and the fish shoal is oxygen deficient at the moment t+1 can be obtained, and when the probability is larger than the threshold value 0.9, the fish shoal is considered to be oxygen deficient.
According to the embodiment, various abnormal probabilities are calculated and compared with the abnormal threshold, so that the association application of the implicit parameter and the explicit parameter is realized, the scientific prediction of the abnormal state is realized, and the technical effect of accurate abnormal monitoring is achieved.
5) Step S5: if the abnormality occurs, executing the cultivation measures corresponding to the abnormality to correct the abnormality:
and according to the result of the step S4, if the abnormal state is detected, the intelligent control is used for executing the cultivation measures. Taking abnormal oxygen deficiency of the fish shoal as an example, controlling the oxygenation equipment to perform oxygenation.
The embodiment has the following technical advantages:
compared with common aquaculture monitoring, the method provided by the invention has the advantages that the hidden parameters related to aquaculture and the relation between the hidden parameters and the dominant parameters are considered, the defect that the monitoring parameters are only the dominant parameters in the traditional method is overcome, meanwhile, for the prediction of abnormal states, the prediction and control are more intelligent and scientific through multi-parameter monitoring of the hidden parameters and the dominant parameters, and the technical effects of comprehensive monitoring, real-time monitoring and accurate abnormal monitoring are achieved.
Example 2
Referring to fig. 3, the present embodiment provides an aquaculture state monitoring device based on the hidden markov model of embodiment 1, including:
the model building module is configured to build an aquaculture model based on a hidden Markov model according to the relationship between the hidden parameters extracted by the culture rule and the culture dominant state, wherein the hidden parameters are used for representing the culture abnormal state information;
specifically, the model building module comprises a sample unit, a training unit and a verification unit; the sample unit is configured to construct a fish swarm dominant state matrix according to samples of the aquaculture water environment information and the fish swarm activity information at a plurality of time points, and simultaneously record the fish swarm states at each time point to form a relationship between recessive parameters and the aquaculture dominant states; the training unit is configured to perform hidden Markov model training according to the shoal state and the shoal dominant state matrix so as to obtain an optimal aquaculture model; the verification unit is configured to calculate the parameters of the aquaculture model by adopting a Baum-Welch algorithm and verify the training result of the aquaculture model;
the detection sensing module is configured to acquire culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information;
specifically, the detection sensing module comprises a receiving and processing unit, a water environment sensor and a vision sensor; the receiving and processing unit is configured to receive the water environment data monitored by the water environment sensor to form aquaculture water environment information; the receiving processing unit is further configured to read and scan the culture monitoring pictures at different moments monitored by the visual sensor, compare the culture monitoring pictures with the characteristic values of the fish shoals, and acquire a fish shoal position information matrix at different moments; the receiving processing unit is further configured to calculate and obtain the fish swarm activity information according to the fish swarm position information matrixes at different moments, wherein the fish swarm activity information comprises fish swarm speed, acceleration, depth and dispersion;
the data processing module is configured to acquire an observation sequence according to the culture dominant state;
the data processing module is specifically configured to combine the aquaculture water environment information and the fish school activity information at the same time point according to the aquaculture water environment information and the fish school activity information to form a time-based observation sequence;
the data processing module is further configured to calculate various anomaly probabilities according to the observation sequence and the aquaculture model, and if the anomaly probabilities are larger than the corresponding anomaly thresholds, the corresponding anomalies are considered to occur, and if the anomaly probabilities are not, the anomalies are considered not to occur.
Preferably, the embodiment further includes an exception handling module configured to execute the cultivation measure corresponding to the exception to perform exception correction.
Compared with common aquaculture monitoring, the method has the advantages that the hidden parameters related to aquaculture and the relation between the hidden parameters and the dominant parameters are considered, the defect that the traditional method is only used for monitoring the parameters as the dominant parameters is overcome, meanwhile, for the prediction of abnormal states, the prediction and control are more intelligent and scientific through multi-parameter monitoring of the hidden parameters and the dominant parameters, and the technical effects of comprehensive monitoring, real-time monitoring and accurate abnormal monitoring are achieved.
The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the above embodiments. Even if various changes are made to the present invention, it is within the scope of the appended claims and their equivalents to fall within the scope of the invention.
Claims (8)
1. A method for monitoring aquaculture conditions based on hidden markov models, comprising the steps of:
s1: establishing an aquaculture model based on a hidden Markov model according to the relationship between hidden parameters extracted by the culture rule and the culture dominant state, wherein the hidden parameters are used for representing the information of the culture abnormal state;
s2: acquiring the culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information;
s3: obtaining an observation sequence according to the culture dominant state;
s4: according to the observation sequence and the aquaculture model, calculating to obtain various abnormal probabilities, if the abnormal probabilities are larger than corresponding abnormal thresholds, corresponding abnormalities are considered to occur, and if not, no abnormalities are considered to occur;
in the step S1, a fish swarm dominant state matrix is constructed according to samples of aquaculture water environment information and fish swarm activity information at a plurality of time points, and meanwhile, the fish swarm state at each time point is recorded to form a relationship between the recessive parameters and the aquaculture dominant state;
performing hidden Markov model training according to the shoal state and the shoal dominant state matrix to obtain the optimal aquaculture model;
and calculating the parameters of the aquaculture model by adopting a Baum-Welch algorithm, and verifying the training result of the aquaculture model.
2. The method for monitoring the aquaculture state based on the hidden markov model according to claim 1, wherein in the step S2, the step of obtaining the fish school activity information includes:
reading and scanning culture monitoring pictures at different moments, comparing the culture monitoring pictures with fish swarm characteristic values, and obtaining fish swarm position information matrixes at different moments;
and calculating the shoal of fish activity information according to the shoal of fish position information matrixes at different moments, wherein the shoal of fish activity information comprises the speed, the acceleration, the depth and the dispersion of the shoal of fish.
3. The method for monitoring the aquaculture state based on the hidden markov model according to claim 1, wherein the step S3 specifically comprises the following steps:
and combining the aquaculture water environment information and the fish school activity information at each same time point according to the aquaculture water environment information and the fish school activity information to form the time-based observation sequence.
4. A method of monitoring the state of aquaculture based on hidden markov models according to any one of claims 1 to 3 wherein step S4 is followed by step S5:
if the abnormality occurs, the cultivation measure corresponding to the abnormality is executed to correct the abnormality.
5. An aquaculture state monitoring device based on hidden markov model, characterized by comprising:
the model building module is configured to build an aquaculture model based on a hidden Markov model according to the relationship between the hidden parameters extracted by the culture rule and the culture dominant state;
the detection sensing module is configured to acquire the culture dominant states at different moments, wherein the culture dominant states comprise culture water environment information and fish swarm activity information;
the data processing module is configured to acquire an observation sequence according to the culture dominant state;
the data processing module is further configured to calculate various abnormal probabilities according to the observation sequence and the aquaculture model, if the abnormal probabilities are larger than corresponding abnormal thresholds, corresponding abnormality is considered to occur, and if not, abnormality is considered not to occur;
the model building module comprises a sample unit, a training unit and a verification unit;
the sample unit is configured to construct a fish swarm dominant state matrix according to samples of the aquaculture water environment information and the fish swarm activity information at a plurality of time points, and record fish swarm states at each time point at the same time so as to form a relationship between recessive parameters and the aquaculture dominant states;
the training unit is configured to perform hidden Markov model training according to the shoal state and the shoal dominant state matrix so as to obtain the optimal aquaculture model;
the verification unit is configured to calculate the aquaculture model parameters by using a Baum-Welch algorithm and verify training results of the aquaculture model.
6. The hidden markov model based aquaculture state monitoring device of claim 5 wherein the detection sensing module comprises a receiving processing unit, an aqueous environment sensor, a visual sensor;
the receiving and processing unit is configured to receive the water environment data monitored by the water environment sensor to form the aquaculture water environment information;
the receiving processing unit is further configured to read and scan the culture monitoring pictures at different moments monitored by the visual sensor, compare the culture monitoring pictures with the characteristic values of the fish shoals, and acquire a fish shoal position information matrix at different moments;
the receiving processing unit is further configured to calculate the shoal of fish activity information according to the shoal of fish position information matrixes at different moments, wherein the shoal of fish activity information comprises a shoal of fish speed, acceleration, depth and dispersion.
7. The hidden markov model based aquaculture state monitoring device of claim 5 wherein the data processing module is specifically configured to combine the aquaculture water environment information and the fish school activity information for each same point in time to form the time based observation sequence according to the aquaculture water environment information and the fish school activity information.
8. The hidden markov model based aquaculture status monitoring device of any one of claims 5 to 7, further comprising an anomaly handling module configured to perform anomaly-corresponding farming measures for anomaly correction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910153100.0A CN109886233B (en) | 2019-02-28 | 2019-02-28 | Aquaculture state monitoring method and device based on hidden Markov model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910153100.0A CN109886233B (en) | 2019-02-28 | 2019-02-28 | Aquaculture state monitoring method and device based on hidden Markov model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109886233A CN109886233A (en) | 2019-06-14 |
| CN109886233B true CN109886233B (en) | 2023-05-09 |
Family
ID=66930017
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910153100.0A Active CN109886233B (en) | 2019-02-28 | 2019-02-28 | Aquaculture state monitoring method and device based on hidden Markov model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109886233B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111325522B (en) * | 2020-02-24 | 2022-12-06 | 浪潮通信信息系统有限公司 | Intelligent fishery breeding management method and system based on Internet of things |
| CN112841052B (en) * | 2020-12-30 | 2023-03-31 | 中农御品(厦门)智能设备有限公司 | Intelligent pigsty monitoring system |
| CN117473446B (en) * | 2023-12-27 | 2024-03-15 | 深圳市明心数智科技有限公司 | Anomaly detection and processing method and system based on deep learning |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102549578B (en) * | 2010-10-01 | 2016-03-09 | 松下电器产业株式会社 | Sorter and sorting technique |
| CN106259095B (en) * | 2016-08-16 | 2019-04-02 | 常州工学院 | A kind of judgment method of the behavior of pond cultured freshwater fish to fish culture state |
| CN107256608A (en) * | 2017-08-14 | 2017-10-17 | 合肥裕亿农业技术开发有限公司 | A kind of cultivation safety-protection system that can be remotely captured based on mobile Internet |
| CN109284921B (en) * | 2018-09-17 | 2021-08-24 | 北京工商大学 | Agricultural irrigation water quality dynamic evaluation method based on hidden Markov model |
-
2019
- 2019-02-28 CN CN201910153100.0A patent/CN109886233B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109886233A (en) | 2019-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109886233B (en) | Aquaculture state monitoring method and device based on hidden Markov model | |
| CN112200011B (en) | Aeration tank state detection method, system, electronic equipment and storage medium | |
| CN102124964B (en) | Intelligent management system for mariculture | |
| CN115861721B (en) | Livestock and poultry breeding spraying equipment state identification method based on image data | |
| CN114549469A (en) | Deep neural network medical image diagnosis method based on confidence degree calibration | |
| CN115471645A (en) | Knowledge distillation anomaly detection method based on U-shaped student network | |
| CN111985670A (en) | Aquaculture monitoring method and device and storage medium | |
| CN114092378A (en) | Animal health detection method, device, equipment and storage medium | |
| CN112861666A (en) | Chicken flock counting method based on deep learning and application | |
| CN116363494B (en) | Fish quantity monitoring and migration tracking method and system | |
| CN116484287B (en) | Soil moisture prediction method, system, terminal and medium based on error correction | |
| CN110956178B (en) | Plant growth measuring method and system based on image similarity calculation and electronic equipment | |
| CN116431621A (en) | Multi-parameter data processing method for aquaculture water quality | |
| CN108845075B (en) | Compost maturity real-time prediction method based on deep learning network | |
| CN112380486B (en) | Dynamic monitoring and evaluating method and system for sea water ornamental fish culture state | |
| CN110633679B (en) | Automatic pointer instrument indicating identification method and system based on genetic algorithm | |
| CN114022831A (en) | Binocular vision-based livestock body condition monitoring method and system | |
| CN112215107A (en) | Pig behavior identification method and device, electronic equipment and storage medium | |
| CN115272943B (en) | Livestock and poultry feeding abnormity identification method based on data processing | |
| Palconit et al. | Fish stereo matching using modified k-dimensional tree nearest neighbor search | |
| CN111950396B (en) | Meter reading neural network identification method | |
| CN115690585B (en) | Method and system for extracting wheat tillering number based on digital photo | |
| CN114255203B (en) | Fry quantity estimation method and system | |
| CN116645732B (en) | Site dangerous activity early warning method and system based on computer vision | |
| CN116543345A (en) | Generation method of traceable file based on artificial intelligence crop production process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |